Discussion

A study of 1265 lunches eaten at school by adolescents in the Cambridgeshire area has shown that a majority of lunches consumed did not meet the recommendations for intakes of energy and most nutrients at lunch. There were some significant differences between packed lunches and those provided by schools but of concern is the finding that school lunches more frequently exceeded recommended intakes for Na and saturated fat. Lunches eaten at school contributed less than one-third of total energy and nutrient intakes but, for girls in particular, packed lunches contributed less of most nutrients to their daily total intake than school lunches. However as there were no significant differences, with the exception of vitamin C (results not shown), in total daily nutrient intakes between girls eating packed and school lunches, it would appear that nutrient intake at midday was balanced out by that consumed during the rest of the day. However, boys and girls who did not record eating lunch had significantly lower intakes of total daily energy (P < 0·001; results not shown).

Energy and nutrient intakes for boys and girls reflected food choices offered by school meals and packed lunches. School meals provided more folate through a greater provision of vegetables, compared with packed lunches which provided greater amounts of Ca through cheese and yoghurts.

There have been very few reports of midday meals eaten at school by adolescents. The advantage of the ROOTS study is that it comprised a large sample of teenagers from urban and rural schools in the Cambridgeshire area. The 4 d diet diary of estimated portion sizes is one of the best ways to accurately assess dietary intake⁽Reference Bingham, Cassidy and Cole¹¹⁾. The physical assessments included in the study required close contact between the investigators and the schools and the participants, thus improving the motivation of the participants and the quality of the data collected.

The limitation of the present study was that the cohort is not a true national representation of the population as the participants are all from the Cambridgeshire area, although both urban and rural areas were included. Also a high proportion of the participants were from families of high socio-economic status. The National Diet and Nutrition Survey (NDNS) of young people aged 4–18 years in 1997 was more nationally representative and showed that regional differences in energy and macronutrient intakes were very small, although there were some regional differences in micronutrient intakes⁽Reference Gregory and Lowe¹²⁾. Only two weekday diet records per participant were analysed in the present study whereas the CWT guidelines should be applied to the average lunch intake over 5 d. However, evidence from the NDNS data of young people indicated that daily intakes of nutrients are quite similar on weekdays⁽Reference Thane and Stephen¹³⁾. The present 2005–2007 study may not represent the quality of recent school meals as the new nutritional standards were released in 2007 and had not reached secondary schools by the time the data were collected⁽⁶⁾.

The few significant differences between school meals and packed lunches did not, on balance, show either option to provide a healthier diet at lunch. Whereas school meals provided more Fe and folate, they were also higher in saturated fat and Na. There were no biomarkers measured in the present study, but the nutrient intakes would seem to indicate potentially a different outcome from that reported in a comparison of school and packed lunches in relation to markers of cardiovascular health in 13–16-year-olds. That study showed small but significant differences in several cardiovascular risk factors that were potentially more favourable in teenagers who ate school lunches⁽Reference Whincup, Owen and Sattar¹⁴⁾. However, serum folate concentrations were lower in teenagers eating school lunches. Those data were collected earlier (1998–2000) as part of the Ten Towns Heart Health study so the participant profiles may have been rather different.

The types of foods provided by school or packed lunches, and hence the nutrients contained in those meals, reflected the meals’ suitability to either kitchen preparation or transport to school. Protein was significantly higher in school meals for boys, but not girls, due to a greater consumption of meat and pulses which also contributed to the marginally higher intakes of Fe. With the exception of bread there was less opportunity to provide Fe in packed lunches due to the types of foods which lend themselves to packed lunch; cooked meats are a possibility but are relatively expensive. The food-based standards⁽³⁾ recommended that schools serve three servings of red meat per week and, although there was no significant difference in total Fe intake between packed lunches and school lunches, the latter provided a small but significantly greater quantity of haem Fe (results not shown); as a better absorbed form of Fe, haem Fe is of particular importance to adolescent girls. Significantly higher intakes of folate were also observed among those eating school meals, which was due to their greater content of pulses and vegetables. The food-based standards⁽³⁾ included the provision of at least two portions of fruit and vegetables per child per day. However, the mean consumption of vegetables in school meals was only 22·3 g and only 29 % of participants were consumers. Vegetables were the main source of vitamin C for pupils eating school meals, as only 16 % of participants consumed fruit in a school meal. Although fruit should have been provided at the schools, options such as biscuits, cake and yoghurts appeared to be more common. Packed lunches, on the other hand, provided most of the vitamin C from fruit and fruit juices, probably due to convenience in transporting these to school. Ca intakes were higher in those consuming packed lunches compared with the school meals as packed lunches were more likely to include bread, cheese and yoghurt. Intakes of total sugars, provided by confectionery, fruit, fruit juices, fizzy drinks and squashes, were greater in those consuming packed lunches. Both types of lunch failed to meet the CWT recommendations for energy, NSP and many micronutrients⁽Reference Crawley²⁾. The mean intakes of some nutrients were misleading as the range was very wide with the medians consistently lower than the means. This resulted in a very high percentage of individuals failing to meet the standards. Vitamin C intakes were particularly skewed; the mean intakes for all lunches were above the CWT standard but the medians were all below, and 68 % of all participants eating school lunches fell below the recommended intake of 14 mg.

A further analysis of the data from the NDNS of young people aged 4–18 years in 1997 also compared the percentage contribution of school lunches to total energy and nutrient intakes⁽Reference Nelson, Lowes and Hwang¹⁵⁾. No data were reported for packed lunches and data were presented for all secondary-school pupils up to the age of 18 years, whereas the ROOTS data were from 14–15-year-olds only. The contribution of secondary-school lunches to total nutrient intake (23–30 %) in the NDNS was very similar to that in the ROOTS study except for fat, sugar and vitamin C, of which there was a slightly greater contribution from school lunches to the total daily intake. Actual intakes of energy and nutrients at lunchtime were not reported but the total daily intakes in the NDNS sample were very close to those of the ROOTS sample except for lower Ca and vitamin C intakes. This may be due to the NDNS being a nationally representative sample with a wider range of socio-economic backgrounds whereas the ROOTS cohort profile had a greater proportion of the highest socio-economic group compared with the total UK⁽Reference Goodyer, Croudace and Dunn⁹⁾. Also the NDNS data date from 1997 and there has been an increase in vitamin C intakes in all ages over this period, particularly with the increase in fruit juice consumption⁽Reference Bates, Lennox and Swan¹⁶⁾. In common with the ROOTS study, mean intakes of energy, fibre, Ca, Fe and folate of those eating school lunches were all below the CWT guidelines while the Na intake was above. This comparison shows that in the 10 years between the NDNS study and the ROOTS study there has been little improvement in the quality of school lunches or, possibly, in the choices made by pupils.

The 2004 report by Nelson et al. for the DfES showed the results of 5695 school meals consumed in seventy-nine secondary schools⁽Reference Nelson, Bradbury and Poulter⁴⁾; this report followed the 2001 introduction of the food-based guidelines⁽³⁾. Although the energy, macronutrient and Na contents of these meals were greater than the school lunches of the ROOTS study, the contents of Fe, Ca and vitamin C were very similar. The percentages of meals failing to reach the CWT guidelines for Fe and Ca were also similar. In the ROOTS study 94 % and 83 % of school lunches failed to meet the guidelines for Fe and Ca, respectively, compared with 93 % and 80 % of the 814 school meals in Nelson et al.'s report⁽Reference Nelson, Bradbury and Poulter⁴⁾.

There are very few data for comparison of packed lunches eaten by secondary-school pupils. School lunches were compared with packed lunches consumed by 621 children aged 7 years in South West England⁽Reference Rogers, Ness and Hebditch⁸⁾. The results showed that both types of lunch needed improvement but the standard of lunches brought from home was worse than that served at school. A study solely of packed lunches consumed by 1294 children aged 8–9 years showed that only 1·1 % met the school meal standards⁽Reference Evans, Greenwood and Thomas¹⁷⁾. A survey published by the University of Plymouth on the diets of primary-school children found school meals to be healthier than packed lunches, as the children taking packed lunches consumed approximately double the amount of sugar and 50 % more Na and saturated fat, compared with those having a school lunch⁽Reference Rees, Richards and Gregory⁷⁾. Comparison of this study with lunches brought to school by teenagers is not really valid as the latter would probably had more freedom to choose the content of their lunches and may even have bought food on the way to school. In fact the ROOTS study showed little difference in the nutrient content of packed lunches compared with school lunches; rather the differences were in the types of foods consumed. The lack of vegetables in packed lunches and the inclusion of soft drinks, savoury snacks and confectionery are indications that food-based guidelines for school lunches were not carried over into recommendations for packed lunches. A recent report from Ofsted has confirmed that, despite government initiatives, the number of pupils eating school meals has declined and more pupils are bringing in packed lunches. The report raised concerns that the healthier school meals were proving too expensive for low-income families who did not qualify for free school meals. They also reported that not all schools had policies on packed lunches but those that did tended to focus on what should not be included rather than giving guidance on how to provide a balanced meal in a cost-effective way⁽¹⁸⁾.

It is a cause for concern that the midday meal eaten at school by the ROOTS teenagers, while contributing between a quarter and a third of daily energy and nutrients, did not meet the CWT guidelines and, in addition, a significant number of pupils did not report any midday meal. This is an important age group in terms of their increased requirements for some nutrients. Teenage boys have the greatest requirement for energy to match their rapid growth during these years, while teenage girls have increased Fe requirements on reaching puberty. Intakes of Fe and Ca were most likely to fail to meet the guidelines. Median intakes of vitamin C and carotene were very low, indicating that fruit and vegetable intake of some participants could not have reached the food-based guideline: ‘Not less than two servings per day per child must be provided; at least one should be vegetables or salad and at least one should be fruit’⁽¹⁹⁾. Consumption of fruit and vegetables has been shown to have an inverse relationship to several chronic diseases, such as diabetes⁽Reference Bates, Lean and Mansoor^20, Reference Harding, Wareham and Bingham²¹⁾ and CVD⁽Reference Bazzano, Serdula and Liu²²⁾. It has also been shown that increased consumption of fruit and vegetables was positively associated with bone mineral content in another group of Cambridge adolescents who were at an age when they should have been maximising their peak bone mass to ensure bone health in later life⁽Reference Prynne, Mishra and O'Connell²³⁾. While being very vulnerable in terms of many aspects of health, teenagers are the age group most difficult to reach. They are more independent, having freedom to make both good and bad choices and are also influenced by peer pressure, particularly with an emphasis on body shape. The rolling programme of improvements in school meals had only begun to reach secondary schools in September 2006. New nutrient-based standards and new food-based standards for school lunches were to be adopted by all secondary schools by September 2009⁽Reference Boaden and Thomas²⁴⁾; hence these new guidelines would not have reached secondary schools by the time the data were collected for the present study. However, at the time when the present survey was starting to be carried out there was a very popular television programme that focused on the need to improve school meals with much subsequent newspaper coverage. This should have raised awareness of the importance of healthy school meals both in those responsible for choosing the contents of packed lunches and those in charge of school catering. The latter already had the food-based guidelines in place, although there was evidence that these were not always adhered to⁽Reference Nelson, Bradbury and Poulter⁴⁾. One might also expect to have seen a change in pupils’ attitudes that could have been reflected in their choice of foods. However, the recent analysis of the first year of the new NDNS rolling programme (2008–2009) indicates that inadequate intakes of some nutrients, particularly Fe, Zn, K and Mg, continue to be a problem in teenagers⁽Reference Whitton, Nicholson and Roberts²⁵⁾. For girls aged 11–18 years, the mean Fe intake was 58 % of the RNI and 46 % had intakes below the lower RNI. Only 7 % of girls and 22 % of boys aged 11–18 met the 5-a-day fruit and vegetable guideline.

The provision of healthy school lunches at a reasonable cost provides a perfect opportunity not only to improve the health of schoolchildren but also to educate them. What they learn about nutrition and health in the school setting may percolate out into their home environment and further. As it is recognised that many children will not take up the offer of school lunches, the present study shows that it is equally important that attention is given to the components of packed lunches and relevant advice offered to parents and children.